CONDUCTANCE AND ACTIVITY COEFFICIENTS OF AQUEOUS RAREEARTH HALIDES 47; 1
Oct. 5 , 1932
TABLE I
Solute
Solvent
CioHs
(C4Fg)zO
CioHs
(C3F7)sN
C ~ H T N O ? (C4F9)?0 C7H7K02
(C3F7)aX
C2Cla
(C4Fe)20
C2C16
(C3F7)3N
SOLUBILITIES I N FLUOROCARBON DERIVATIVES Solubility, Solubility, Solubility, g./lOO, g. ;olvent mole fraction g /lo0 g. s p e n t a t 25 a t 26' a t 35 Measured Average Measured Ideal Measured Average
0.0726 .0733 .0774 .0723 .0723 .OS18 .OS06 .0850 ,0850 ,474 ,475 ,516 .509 ,525 .480
0.0730
0.00257
0.299
.0740
.00300
,299
,0812
.00268
,571
.OS50
.00323
,571
,474
.00901
.0526
.508
.0111
.0526
fraction and at 35" 0.304. This is much closer to the ideal figure despite the fact that dibutyl ether has a dipole moment whereas neither naphthalene or (C4F&O have an appreciable moment. From this, one might expect naphthalene to be more ideally soluble in (C4Fg)20 than in (C4He)20. The reverse is found. This can be explained using the concept of interpenetration.2 The solubility of nitrobenzene falls far below the ideal in both fluorocarbon derivatives. This is due primarily to interpenetration but also to the fact that it is a dipolar material, differing in this respect from the fluorocarbon derivatives. Hexachloroethane comes much more nearly being ideally soluble in the fluorocarbon derivatives, probably because of the lack of hydrogen atoms in [CONTRIBUTION
197, INSTITUTE
FOR
Solubility, mole fraction a t 35' Measured Ideal
0.116 .112 .115 .113
0.114
0.00402
0.384
-114
.00481
.384
,123 .119 .134 .140 .781 .780 .743 .741
,121
.00399
.712
.137
,00517
.712
.78t
,0145
.0709
.742
.OlfiO
.0709
the molecule resulting in little interpenetration between the solute molecules. From the difference in solubility a t two temperatures, the heat of solutions was calculated. For naphthalene in (C4Fs)20 this was 8150 cal. per mole and in (CsF7)aN was 7850. This is compared with the ideal value of 4580. For p-nitrotoluene in (C4Fg)20 this was 7320 and in (C~F,)SN was 8620. This is compared with the ideal value of 4010. The excess of these values above the ideal gives a rough measure of the heat of interpenetration. As can be seen from the figures, there appears to be little difference in solubility of these substances in the fluorocarbon oxide or fluorocarbon nitride. GAINESVILLE, FLORIDA
ATOMICRESEARCH AND THE DEPARTMENT OF CHEMISTRY,
IOWA STATE COLLEGE]
Conductances, Transference Numbers and Activity Coefficients of Aqueous Solutions of Some Rare Earth Halides at 2 5 0 1 BY F. H. SPEDDING AND I. S. Y A F F E ~ RECEIVED M A Y 24, 1952 The equivalent conductances, cation transference numbers and activity coefficients a t 25" of aqueous solutions of LaBra, PrBrs, NdBrs, GdBrg, HoBrs, ErBrs and GdCls have been determined for concentrations up to 0.1 N . Also the equivalent conductances of NdClr have been redetermined. The Onsager law for conductance is obeyed for all halides studied up to a concentration of approximately 0.001 N . The transference numbers, determined by the moving boundary method,were found t o be linear functions of the square root of the concentration; the slopes of these functions differed by a factor of approximately one-fifth from those predicted by the Onsager law. The activity coefficients, determined by the use of concentration cells with transference, were found to agree with those predicted by the Debye-Huckel law providing that the experimentally determined &. values were used to calculate the theoretical coefficients. All data reported are correlated with data on the rare earth chlorides previously reported.
Introduction This paper is the fourth in a series concerning the electrolytic behavior of aqueous solutions of rare earth compounds. The earlier paper^^-^ have (1) Work was performed in the Ames Laboratory of the Atomic Energy Commission. (2) Taken in part from the Ph.D. thesis submitted to the graduate faculty of Iowa State College. 1952. (3) F. H. Spedding, P. E. Porter and J . M. Wright, THISJOURNAL, 74, 2055 (1952). (4) F. H. Spedding, P E. Porter and J. M . Wright, ibid., 74, 2778 (1952).
(5) F. H. Spedding, P. E . Porter and J. M. Wright, ibid., 1 4 , 2781
(1952).
presented data on the conductances, transference numbers and activity coefficients of several rare earth chlorides. This paper extends this investigation to other rare earth halides, particularly the bromides, and attempts to correlate the data between the two series of halides. As discussed in the first article of this s e r i e ~such , ~ information should be of considerable value in the study of the various factors which enter into the modern theories of electrolytic behavior. Experimental Except as specifically discussed below, the experimental
F. H. SPEDDING AND 1. s. Ir.4FFE
4752
VOl. '73
TABLE I EQUIVALENT COXDUCTANCES OF AQUEOUS SOLUTIONS OF SONE RAREEARTH HALIDESAT 25 Normality
Lanra
SdBrz
PrBri
0.0000
147.6 147.6 143. 6 143,C, 142.0 142.0 .0007 140.2 140.2 138.7 138.7 .0010 .0020 137.8 137.7 .0040 131.8 131.7 .0070 127.7 127.6 124.5 .OlOO 124.6 ,0200 118.5 118.4 112.2 112.1 .0400 .0700 109 0 108.9 103.8 103.6 ,1000 0 The data of this table are smoothed them from t.he authors upon request .0002 .0004
I
GdRri
Equivalent conductances, mhos ciii. -1 147.t; JITi.6 144.4 143. f j 111.7 140.5 142.0 140.0 138.8 140.2 138.2 137.0 138.7 136.8 135.ti 137.7 133.0 132.5 131.6 128.8 128.4 127.6 125 0 124.5 124.4 122.2 121,r, 118.3 116.1 115.8 112.0 110.1 109.3 108.8 105,S 104.2 103.4 102.0 101.0 values. Anyone desiring the exact d a t a
procedures and apparatus were the same as those previously rep0rted.~-6 All measurements were made at 25 2~ 0.02'. The samples of La203,PrsOlt and NdlO3 used for this work had the same analyses as those previously reported.3 The ourities of the other rare earth oxides used were: GdlOq: '< 0.1% SmzOs, 0.025% Tb40-i, faint trace of Cad; Ho203: 0.05% Dy?03, 0.02% Y ~ 0 3 , 0 01% ErzO?, faint trace of CaO; and Er203: 0.05% YbAOl, 0 05% Tm2O3,0.05% HozO3,0.2% YZ03, no DyZOadetected, faint trace of CaO. All analyses were made by spectrographic methods6 except t h a t for T b r 0 7in Gd?Oo which was made by a new fluorescimetric technique.? A11 oxides were further purified by a minimum of two oxalate precipitations from solutions containing a slight excess of the hpdrohalide acid with which they were finally dissolved. GdCl, solutions were prepared from the anhydrous halide.3 Attempts were made to prepare anhydrous rare earth bromides by a n analogous method, using purified dry HBr gas. However, although solutions obtained from the dried salts were clear, their pH values ranged from 7.10 to 7.30, and analyses for the rare earth and for the bromide ions indicated that the bromide content in all cases was less than the required stoichiometric amount. Therefore the following method was employed to prepare the rare earth bromide solutions. The purified oxide was dissolved in a n insufficient quantity of redistilled hydrobromic acid a t room temperature. n'hen the reaction was completed, the undissolved oxide was filtered from the solution, and the filtrate was diluted t o a predetermined volume with conductivity water. Analyses of these solutions indicated that the two ionic species were present in the proper stoichiometric ratio t o within O.I$&. After filtering the PrBr3 solution, the filtrate had to be boiled to remove the free Br2 formed when dissolved. Dilutions ns using recalibrated esired concentrations. paration, solutions of SdC13were preparcd by this method a n d were also prepared from the anhydrous chloride. TI equivalent conductances of these solutions fell on the IIK smooth curve and both sets of dnta eutrapn1;rtetl t o tt .:line w l u e a t infinite dilution. LiBr was used as the indicator solution for the determinations of the cation transference numbers of the rare earth hornides and a Ag,AgBr electrode was used as the cathode. The LiBr stock solution mas prepared by a method analogous to that used for the preparation of LiC1.8 The transference numbers of LiBr were calculated from transference numher data on KBrg and 1,iClIO and from the conductance data
- -
-
(6) V. A. Fassel and 11. .A. U'ilhelm, J . Oplicnl Sac. A m . . 38, 518 V. .4.Fassel. i b i d . . 39, 187 (1949). (7) V. A . I'assel. R. 11. ITeidrl and F. Huke, A n a l . C i i e m . , 94, 606 (19.52) ( 8 ) G. Pcatchll-ii i ~ ~ ~ h. r ! S. I'rentis~. T I ~ I S J O I J X N A L , 66, 4355
(1948):
(1933) ((9) I,. G Longsworth, i b i d , 67, 118.7 ( 1 9 3 5 ) . i10) I, (> I.ongsi\orth. r h r i ! , 6 4 . 9711 1 1 ' ) 3 ?
IIoBrr
ErBri
Sd Clt
OR
MC13
144,o 145.8 143.7 140.1 111.8 139.8 138.4 140.2 138.1 138.4 136.3 1.36,7 135.2 136.9 134.9 132.2 133.3 131.3 128.2 128.7 127.1 124.2 124.6 123.0 121.3 121.5 120.2 115.0 114.0 115.6 107.7 107.0 109.0 103.9 101.7 101.5 100.7 98.6 98.4 for the experimental points may obtain
on KBr" and on LiCl.l* As a n approximate check upon the calculated transference numbers of LiBr, the transference numbers of K B r were determined at several concentrations, using LiBr as the indicator solution. Within experimental error, the transference numbers of K B r so obtained agreed with those of Longsworthe who used the autogenic boundary technique. Also, in view of the close correlation between experiment and theory for alkali halides, i t is believed that the calculated transference numbers of LiBr are in error by no more than 5%. The transference numbers of indicator solutions need not be known t o greater accuracy. Since infrared heating was used for the constant temperature bath (A. H. Thomas Catalog No. 9926-D) in which the concentration cells with the transference were immersed, i t was thought expedient to paint the outsides of these cells black t o protect the Ag,AgBr electrodes from possible photodecomposition.
Results
A. Conductances.-The equivalent conductances of the eight rare earth halides investigated are listed in Table I. Within experimental error, the conductances of LaBr3, PrBra and NdBr3 are identical over the concentration range studied although small differences are observed in the higher concentrations. For the extrapolation to infinite dilution, Xo for the chloride ion was taken as 76.34 mhos cm.-l, and that for the bromide ion as 75.15 mhos cm.-l. The latter value was obtained by averaging the values of A0 for KBr obtained by various investigators,l1,l3 and subtracting from that average the accepted value of the limiting conductance of potassium ion, 73.50 mhos cm.-'. The extrapolation curves are illustrated in Fig. 1. The horizontal portion of each curve indicates the concentration range in which the Onsager law is obeyed, up to approximately 0.001 hTfor each salt. Since the conductances of NdC13 previously reported3 were believed to be slightly less accurate than the other salts reported, they were redetermined. The new values are listed in Table I. The equivalent conductances a t infinite dilution, A,,, for the rare earth halides reported here and for (11) G. Jones and C. F. Bickford, ibid., 66, GO2 (1934). (12) T. Shedlovsky. ibid., 64, 1411 (1932); D. A. MacInnes, T. Shedlovsky and L. G. Longsworth, ibid., 64, 2758 (1932). (13) G . C. Benson and A R . Gordon, J . Chem. P h y s . , 13,473 (1945); 1%. Zeldes, Dissertation, Yale University, 1947, cited by Warned and Owen, "The Physical Chemistry of Electrolytic Solutions," 2nd F d i ~ t o i i .Reinhold P u t j l . C a r p , , New Y o r k , S V , 1) . W l , 1'1.50
OCt. 5 , 1952
152
-
La Era
O
A
PrEr3-E NdBr, -E Gd Bra -E
M
HoBr,-P
E
Er Bra-=
M
Nd GI,-=
6
GdCI,-m
0 R E. Br, 0 0
t
\
00
-
b
a 144-
W
-
2 2
143-
0
-
If, I41 -
0 3
6-99
0
142-
0
-
0,O-THIS LABORATORY
9 - JONES 8 EICKFORD (3- MACINNES 8 SHEDLOVSKY
JJ s
-
140
Lo Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb IJJ 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
I
0
CI,
145-
148-q-
142
-
W
-
144
0 R. E
0
150 0
4753
OF AdQUEOUSR.IRE E.\RTHH.4LIDES
ACTIVITYCOEFFICIENTS
-I
x 0
I54
CONDUCTANCE AND
.
0.02
1
1
.
1
.
1
004 006 090 ( NORMALITY 1 P.
1
1
0.10
.
1
012
Fig. 1.-Extrapolation of equivalent conductances t o infinite dilution of aqueous solutions of some rare earth halides a t 25'.
those previously rep0rted8,~~a'~ are plotted in Fig. 2 as a function of the atomic number of the rare earth element. There should be a constant difference between the curves for the two anions of 1.8 mhos; within experimental error, this constant difference is obtained. The only other recent conductance data on rare earth halidesl5 appear t o differ somewhat from our results. However due to the care taken in our measurements and due t o their internal self-consistency, we believe our results are the more nearly correct. Explanations for the variation of the equivalent conductances of the rare earth halides with atomic number have been previously presented. Recent crystallographic data on Ndz(S04j343Hz01s confirms the prediction that Nd(II1) has only one stable coordination number for oxygen a t 2 5 O , namely, nine. B. Transference Numbers.-The cation transference numbers for LaBr3, PrBra, NdBr3, GdBr3, HoBr3, ErBr3 and GdC13 are listed in Table 11. These data were subjected to a least squares analysis and the equations so obtained are tabulated in Table 111. The data previously obtained on the transference numbers of the rare earth chlorides4 were similarly treated and are listed for comparison. The limiting transference numbers, ob(14) L. G. Longsworth and D. A. MacInnes, THISJOURNAL, 6 0 , 3070 (1938). (15) G. Jantsch, H. Grabitsch and E. Lischka, 2. Elckfrochcm., 43, 243 (1937). (16) R . E. Rundle and D. R . Fitzwater, unpublished work.
ATOMIC
NUMBER
Fig. 2.-Equivalent conductances a t infinite dilution for aqueous solutions of rare earth halides at 25'.
tained from conductance data, are also listed in this table along with the limiting slopes as predicted from the Onsager limiting law. Although the intercepts of the least squares lines are fairly close to the limiting transference numbers, the slopes of these lines differ from the theoretical slopes by a factor of approximately one-fifth. This discrepancy between theory and experiment has been previously r e ~ o r t e d , ~but , ' ~as yet there is no adequate explanation for this behavior. The transference numbers of LaBrs, PrBr3 and NdBrs are quite close to each other, with those of neodymium being slightly higher. In view of the close agreement among the conductances of these three salts, the similarity among their transference numbers was expected. The transference numbers of the heavier rare earth bromides decreased with increasing atomic number as did their conductances. This same general trend has been previously reported for the rare earth chloride^.^ From the data in Table 111, i t can be seen that, except for the higher mobility of the bromide ion over that of the chloride ion, the effect of these anions upon the rare earth cations is relatively the same. Erbium appears to be an exception. Before experimentally determining the cation transference numbers of HoBrs, these numbers were estimated from the least squares lines of ErBr3 and of GdBr3, as listed in Table 111, assuming, as a first approximation, that there is a linear variation with atomic number between these two salts. Within experimental error, the transference numbers of HoBrl which were experimentally determined agreed with those estimated by this method.
The standard volurrie and solvent corrections were applied to the otmmwl traiisiereiice iiiiiiibers 1 1 1 obtain those listrtl l',~l)lc I I I'hc Iurtiitl
ErBra (I.
i io-1
,08278 .iL%19
.02759
I).1051
.08174 )
.lJ;35113 02335 , ill 168 ,
o . 4 112 ii.Jii;< -u.ooiu ,4114 .4168 ,4163 -- ,0007 ,4238 ,4238 - ,0005 ,4238 ,4328 ,4328 - ,0002 8 ,4398 - ,0001 ,4405 - , l ! 0 l ) l
o.oooo
o.Ji(i:j
.0000
,4156 ,4234
.Oil01
,0001 ,0003
.4327
,OO03
,44117
,4400
(:dCln i).4321 0.4319 -0.0001 0 0000 n.4.'31ej 4R 17 .4:3Fi9 . 4 x 9 - .000:1 .0000 ,436Ii ,4412 .441(i - .(!nl)2 .I)O01 ,441,; 4420 ,4488
,4532 ,4d99
.4488
,4532 ,4599
- , 0001 - .0001 - .00i)O
,4485
. 0001 . 1JlllJ1
. 4532
,0003
,4602
TABLE I11 SUMMARY O F D.4T.4 O K TRANSFEREXCE SUMRERS OF
Salt a
E A R T H EIALIDES I,east-?quares equation 7'. =
T+O
RARE
S(?'+O)b
I,aBr3 0.4672 - ( J , 110 0.4707 -l!,558 PrBr:, ,4697- I20 r 4707 - .:%X SdBrj ,4681)- , 113 r',' 4707 - ,558 GdBrj 4689 - 132 .\,','2 .46.32 - ,581 IHoBr; ,4591 - ,594 ErBr 4573 - .,%I9 LaCl 4773 - ,551) CeCl,r -+;Hi - . 113 .I-'Y ,4778 - ..549 I'rCI. 475x 114 .\.I 4772 - , 5,51 StlCI., ,4744- 1 1 2 . 4787 SinCl 4745 - ,112 . ,47:32 - , Xj:j -1.:u c 1 . - i ~ j . I I I ~.\' 1 ,4iO(i - ,.>I 1 GdCI: ,4687 - , ,)I I ISrC1, 4t;o(l - ,118 .4fi38 - ,591 YhClr ,4629 - . 125 -Y','? . 4611~5 - ,603 \Vith the exception of the chloride data reported in this paper, all of the chlorides listed are from the data of SpedS( T+O) = theoretical limiting ding, Porter and \\.right.i slope as calculated from the Onsager equation for transference numbers. ,
id50
- ,0001
4.Wi
--
0001
f
---
0
4,542
-
4.57I
-
1inoo
r i 4269
molal volumes of the rare earth halides required for the volume corrections were calculated from density measurements made with a 50-ml. pycnometer. I n the concentration range studied, the densities were found to be linear functions of the concentration. These equations are listed in Table IV, along with the partial molal volumes calculated from these- equations. The anticipated variation of p and V with atomic number was obtained.
l!O01 01IO 1
TABLE IV I > R X S I T I E S ASD PARTIAL M O L A L I-OLUMES O F S O M E
I7ARTH HALIDES AT Density equatioo'l Salt
4295
- ,0005
4:36,9
-
nno2
1432
--
r11101
0004
443.5
P =
LaBr? 11.99707 -I- 0.3471C PrBrs .99707 .3525C SdBrs ,99707 .35i2C GdBrs ,99707 .3723C HoBri .!I9707 C .3813C ErBr,$ .00707 .x345c CJdCI j .o:i;ci7 $- ?494(' '1 C = molarity.
+ + + +
RARE
2.5" Partial molal volume, ml.
:3 1 . 60 28,25 26,. 95 24.42 23.47 22.52 14.nn
Oct. 5 , 1932
CONDUCTANCE AND ACTIVITYCOEFFICIENTS OF AQUEOUSRAREEARTH HALIDES 4755
C. Activity Coefficients.-Concentration with transference of the type
method and also by the thermal method of Keston." No difference in the behavior of the two types of bromide electrodes was observed, and both types Ag, AgBr, HoBr3 ( M I i)HoBrt ( i t ! ? ) , AgBr, Ag were used interchangeably. The activity coeffiwere used to determine the activity coefficients. cients obtained are listed in Table V. The method Ag,AgCl electrodes were used for all chloride used for the calculations has been previously desolutions; the method of preparation has been pre- ~ c r i b e d . ~For the integration across the concentraviously d e ~ c r i b e d . ~Ag,AgBr electrodes, used for tion gradient, the transference numbers listed in all bromide solutions, were prepared by the above Table I11 were employed. Table VI lists the experimentally determined 3. values and the average TABLE 1MEAX MOLALACTIVITYCOEFFICIENTSOF RARE EARTH absolute difference between the experimentally determined coefficients and those calculated from AT 25' HALIDESOLUTIOS the Debye-Huckel law using these 8. values. E.m.f., E.m.f., Molnlity mv. y& Molality mx.. y& Within experimental error, the experimentally LaBrt PrBr3 determined activity coefficients obeyed the Debye0.03483 -3.8981 0.4340 0.03535 -1,7260 0.4293 Hiickel law. This same agreement has been .02487 o.ooon ,4676 ,03030 o.oooo ,4457 previously observed for rare earth chloride^.^^'^ .01990 ,009945 ,004971 ,003480 ,002486 ,001988 ,0009942
2.6452 11.0389 19.854 24.576 29.038 32.052 41.873
,4902 ,5624 ,6389 ,6741 ,7097 ,7324 ,7862
,02020 .01010 ,005048 ,003633 ,002019 ,
ooioio
KdBra ~1.03178 -3.3935 0.4403 0.03505 ,02383 0.0000 ,4678 ,02553 .oi587 4.8482 ,5095 ,01702 ,007941 13.4223 ,5833 ,008508 ,6838 ,004255 ,003175 25.254 ,7142 ,002382 29.105 .003403 ,7502 ,001588 34.744 ,002552 . nom939 44.569 ,8087 .001701 .0008507 HoBr? -.5,s327 -3.3429 0.0000 1.6815 12.8016 21.705 28.403 33,805 ,0007943 43.385
n.03975 ,03179 ,02384 ,01589 ,007944 ,003972 ,002383 ,001588
0.4291 ,4496 ,4758 ,5183 ,5907 ,6644 ,7133 ,7531 ,8105
GdCI, -6.7271 0.4172 -8.8901 ,4429 o .oooo ,4787 6.1070 ,5345 11.0586 ,5816 ,6572 19,919 24.645 ,6962 .no1171 36.;ii3 ,7728
0.03515 ,02734 ,01952 ,01171 ,007808 ,003904 ,002733
4.7698 13.0286 21.837 26.613 33.938 43.754
cells
,4861 ,5654 ,6335 ,6667 ,7321 ,7872
GdBr3 -3.2470 0.4229 0,0000 .4523 4.6552 13.1131 .5695 21.818 .6483 24.636 .6759 28.535 .7027 34.073 ,7408 43.752 .SO22
TABLE VI MEANDISTANCESOF CLOSESTAPPROACH FOR RAREEARTH AND THE AVERAGEDIFFERENCESBEHALIDE SOLUTIONS TWEEN CALCULATED AND OBSERVED ACTIVITYCOEFFICIENTS Salt
P.
A"/&
LaBr3 PrBr, NdBrt GdBrt HoBr? ErBr? GdC1,
6.20 6.10 6 06 5 72 6.42 5.90 5.63
n. 0012 ,0026 ,0011 ,0012 ,0008 ,0015 ,0010
Except that the 3. values of the rare earth bromides are larger than those of the corresponding chlorides,j there does not appear to be a regular variation among these values. Unfortunately, the method used to calculate these values is not SUEErBra 0.03693 -3.1028 0,4204 ciently sensitive to determine whether these varia,02769 0.0000 ,4500 tions are real. Undoubtedly, these values are also ,01846 4.62% ,4894 influenced by the approximations inherent in the '.009229 12,6457 ,5665 Debye-Hiickel theory which may not be valid for 004613 21,173 ,6400 these 3-1 electrolytes. Before more definite con,002768 27,538 ,7002 clusions may be made concerning the electrolytic ,001845 32,970 ,7349 behavior of rare earth solutions, data on the rare ,0009225 42.356 ,7949 earth elements not yet investigated should be obtained and correlated with other physical measurements on these elements. Acknowledgments.-The authors wish to thank Dr. E. I. Fulmer for his encouragements and helpful advice during the course of this work, and Mr. Sigmund Jaffe for his assistance in completing the transference number determinations o n GdCl,. ,
AMES,IOVA (17) A. S. Keston, THIS J O U R X A L , 67, 1671 (193:) (18) T.Shedlovsky, ibid , 73, 3880 (19.50).
